Skip to main content

Table 2 Prediction performance of various baseline measurements and their combinations in terms of the AUC Score

From: Sparse learning and stability selection for predicting MCI to AD conversion using baseline ADNI data

 

AUC score

ADAS-cog total 11

0.7024

ADAS-cog total 13

0.7248

ADAS-cog subscore Q1 Word Recall

0.6830

ADAS-cog subscore Q2 Commands

0.1581

ADAS-cog subscore Q3 Construction

0.4899

ADAS-cog subscore Q4 Delayed Word

0.6842

ADAS-cog subscore Q5 Naming

0.3202

ADAS-cog subscore Q6 Ideational

0.5142

ADAS-cog subscore Q7 Orientation

0.4836

ADAS-cog subscore Q8 Word

0.6062

ADAS-cog subscore Q9 Recall

0.2581

ADAS-cog subscore Q10 Spoken

0.3914

ADAS-cog subscore Q11 Word Finding

0.5436

ADAS-cog subscore Q12 Comprehension

0.2756

ADAS-cog subscore Q14 Number Cancellation

0.4451

ADAS-cog subscore Q1-Q14

0.7598

Age

0.5123

Years of Education

0.5090

MMSE Score

0.5916

CDR-SB

0.6064

ADAS total 13 + ADAS subscores

0.7561

ADAS total 13 + ADAS subscores + MMSE + CDR-SB

0.7674

FAQ

0.6874

LDEL: Logical Memory delayed

0.6573

LIMM: Logical Memory immediate

0.6136

MRI (237)

0.7214

Lab tests (18)

0.5348

APOE genotyping

0.5473

TRAA: Trail Making Test: Part A

0.5944

TRAA: Trail Making Test: Part B

0.6187

  1. The numbers in the second column are the leave-one-out AUC score which may be significantly lower than 0.5. The number in the parenthesis denotes the number of measurements involved.